1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
| # NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=tahiti -mattr=+fp32-denormals -run-pass=legalizer %s -o - | FileCheck -check-prefix=F32DENORM %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=tahiti -mattr=-fp32-denormals -run-pass=legalizer %s -o - | FileCheck -check-prefix=F32FLUSH %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=fiji -mattr=+fp32-denormals -run-pass=legalizer %s -o - | FileCheck -check-prefix=F32DENORM %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=fiji -mattr=-fp32-denormals -run-pass=legalizer %s -o - | FileCheck -check-prefix=F32FLUSH %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=fiji -mattr=+fp32-denormals -run-pass=legalizer %s -o - | FileCheck -check-prefix=F32DENORM %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=fiji -mattr=-fp32-denormals -run-pass=legalizer %s -o - | FileCheck -check-prefix=F32FLUSH %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx10 -mattr=+fp32-denormals -run-pass=legalizer %s -o - | FileCheck -check-prefix=F32DENORM %s
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx10 -mattr=-fp32-denormals -run-pass=legalizer %s -o - | FileCheck -check-prefix=F32FLUSH %s
---
name: test_fmad_s32
body: |
bb.0:
liveins: $vgpr0, $vgpr1, $vgpr2
; F32DENORM-LABEL: name: test_fmad_s32
; F32DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; F32DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; F32DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
; F32DENORM: $vgpr0 = COPY %3(s32)
; F32DENORM: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[COPY]], [[COPY1]]
; F32DENORM: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FMUL]], [[COPY2]]
; F32FLUSH-LABEL: name: test_fmad_s32
; F32FLUSH: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; F32FLUSH: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; F32FLUSH: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
; F32FLUSH: [[FMAD:%[0-9]+]]:_(s32) = G_FMAD [[COPY]], [[COPY1]], [[COPY2]]
; F32FLUSH: $vgpr0 = COPY [[FMAD]](s32)
%0:_(s32) = COPY $vgpr0
%1:_(s32) = COPY $vgpr1
%2:_(s32) = COPY $vgpr2
%3:_(s32) = G_FMAD %0, %1, %2
$vgpr0 = COPY %3
...
---
name: test_fmad_s32_flags
body: |
bb.0:
liveins: $vgpr0, $vgpr1, $vgpr2
; F32DENORM-LABEL: name: test_fmad_s32_flags
; F32DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; F32DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; F32DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
; F32DENORM: $vgpr0 = COPY %3(s32)
; F32DENORM: %4:_(s32) = nnan G_FMUL [[COPY]], [[COPY1]]
; F32DENORM: %3:_(s32) = nnan G_FADD %4, [[COPY2]]
; F32FLUSH-LABEL: name: test_fmad_s32_flags
; F32FLUSH: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; F32FLUSH: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
; F32FLUSH: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
; F32FLUSH: %3:_(s32) = nnan G_FMAD [[COPY]], [[COPY1]], [[COPY2]]
; F32FLUSH: $vgpr0 = COPY %3(s32)
%0:_(s32) = COPY $vgpr0
%1:_(s32) = COPY $vgpr1
%2:_(s32) = COPY $vgpr2
%3:_(s32) = nnan G_FMAD %0, %1, %2
$vgpr0 = COPY %3
...
---
name: test_fmad_v2s32
body: |
bb.0:
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $vgpr4_vgpr5
; F32DENORM-LABEL: name: test_fmad_v2s32
; F32DENORM: [[COPY:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr0_vgpr1
; F32DENORM: [[COPY1:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr2_vgpr3
; F32DENORM: [[COPY2:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr4_vgpr5
; F32DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>)
; F32DENORM: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>)
; F32DENORM: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY2]](<2 x s32>)
; F32DENORM: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR %10(s32), %11(s32)
; F32DENORM: $vgpr0_vgpr1 = COPY [[BUILD_VECTOR]](<2 x s32>)
; F32DENORM: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[UV1]], [[UV3]]
; F32DENORM: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FMUL]], [[UV5]]
; F32DENORM: [[FMUL1:%[0-9]+]]:_(s32) = G_FMUL [[UV]], [[UV2]]
; F32DENORM: [[FADD1:%[0-9]+]]:_(s32) = G_FADD [[FMUL1]], [[UV4]]
; F32FLUSH-LABEL: name: test_fmad_v2s32
; F32FLUSH: [[COPY:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr0_vgpr1
; F32FLUSH: [[COPY1:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr2_vgpr3
; F32FLUSH: [[COPY2:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr4_vgpr5
; F32FLUSH: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>)
; F32FLUSH: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<2 x s32>)
; F32FLUSH: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY2]](<2 x s32>)
; F32FLUSH: [[FMAD:%[0-9]+]]:_(s32) = G_FMAD [[UV]], [[UV2]], [[UV4]]
; F32FLUSH: [[FMAD1:%[0-9]+]]:_(s32) = G_FMAD [[UV1]], [[UV3]], [[UV5]]
; F32FLUSH: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[FMAD]](s32), [[FMAD1]](s32)
; F32FLUSH: $vgpr0_vgpr1 = COPY [[BUILD_VECTOR]](<2 x s32>)
%0:_(<2 x s32>) = COPY $vgpr0_vgpr1
%1:_(<2 x s32>) = COPY $vgpr2_vgpr3
%2:_(<2 x s32>) = COPY $vgpr4_vgpr5
%3:_(<2 x s32>) = G_FMAD %0, %1, %2
$vgpr0_vgpr1 = COPY %3
...
---
name: test_fmad_v3s32
body: |
bb.0:
liveins: $vgpr0_vgpr1_vgpr2, $vgpr3_vgpr4_vgpr5, $vgpr6_vgpr7_vgpr8
; F32DENORM-LABEL: name: test_fmad_v3s32
; F32DENORM: [[COPY:%[0-9]+]]:_(<3 x s32>) = COPY $vgpr0_vgpr1_vgpr2
; F32DENORM: [[COPY1:%[0-9]+]]:_(<3 x s32>) = COPY $vgpr3_vgpr4_vgpr5
; F32DENORM: [[COPY2:%[0-9]+]]:_(<3 x s32>) = COPY $vgpr6_vgpr7_vgpr8
; F32DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<3 x s32>)
; F32DENORM: [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<3 x s32>)
; F32DENORM: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY2]](<3 x s32>)
; F32DENORM: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR %13(s32), %14(s32), %15(s32)
; F32DENORM: $vgpr0_vgpr1_vgpr2 = COPY [[BUILD_VECTOR]](<3 x s32>)
; F32DENORM: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[UV2]], [[UV5]]
; F32DENORM: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FMUL]], [[UV8]]
; F32DENORM: [[FMUL1:%[0-9]+]]:_(s32) = G_FMUL [[UV1]], [[UV4]]
; F32DENORM: [[FADD1:%[0-9]+]]:_(s32) = G_FADD [[FMUL1]], [[UV7]]
; F32DENORM: [[FMUL2:%[0-9]+]]:_(s32) = G_FMUL [[UV]], [[UV3]]
; F32DENORM: [[FADD2:%[0-9]+]]:_(s32) = G_FADD [[FMUL2]], [[UV6]]
; F32FLUSH-LABEL: name: test_fmad_v3s32
; F32FLUSH: [[COPY:%[0-9]+]]:_(<3 x s32>) = COPY $vgpr0_vgpr1_vgpr2
; F32FLUSH: [[COPY1:%[0-9]+]]:_(<3 x s32>) = COPY $vgpr3_vgpr4_vgpr5
; F32FLUSH: [[COPY2:%[0-9]+]]:_(<3 x s32>) = COPY $vgpr6_vgpr7_vgpr8
; F32FLUSH: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<3 x s32>)
; F32FLUSH: [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<3 x s32>)
; F32FLUSH: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY2]](<3 x s32>)
; F32FLUSH: [[FMAD:%[0-9]+]]:_(s32) = G_FMAD [[UV]], [[UV3]], [[UV6]]
; F32FLUSH: [[FMAD1:%[0-9]+]]:_(s32) = G_FMAD [[UV1]], [[UV4]], [[UV7]]
; F32FLUSH: [[FMAD2:%[0-9]+]]:_(s32) = G_FMAD [[UV2]], [[UV5]], [[UV8]]
; F32FLUSH: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[FMAD]](s32), [[FMAD1]](s32), [[FMAD2]](s32)
; F32FLUSH: $vgpr0_vgpr1_vgpr2 = COPY [[BUILD_VECTOR]](<3 x s32>)
%0:_(<3 x s32>) = COPY $vgpr0_vgpr1_vgpr2
%1:_(<3 x s32>) = COPY $vgpr3_vgpr4_vgpr5
%2:_(<3 x s32>) = COPY $vgpr6_vgpr7_vgpr8
%3:_(<3 x s32>) = G_FMAD %0, %1, %2
$vgpr0_vgpr1_vgpr2 = COPY %3
...
---
name: test_fmad_v4s32
body: |
bb.0:
liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8_vgpr9_vgpr10_vgpr11
; F32DENORM-LABEL: name: test_fmad_v4s32
; F32DENORM: [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; F32DENORM: [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
; F32DENORM: [[COPY2:%[0-9]+]]:_(<4 x s32>) = COPY $vgpr8_vgpr9_vgpr10_vgpr11
; F32DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<4 x s32>)
; F32DENORM: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<4 x s32>)
; F32DENORM: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY2]](<4 x s32>)
; F32DENORM: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR %16(s32), %17(s32), %18(s32), %19(s32)
; F32DENORM: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[BUILD_VECTOR]](<4 x s32>)
; F32DENORM: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[UV3]], [[UV7]]
; F32DENORM: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FMUL]], [[UV11]]
; F32DENORM: [[FMUL1:%[0-9]+]]:_(s32) = G_FMUL [[UV2]], [[UV6]]
; F32DENORM: [[FADD1:%[0-9]+]]:_(s32) = G_FADD [[FMUL1]], [[UV10]]
; F32DENORM: [[FMUL2:%[0-9]+]]:_(s32) = G_FMUL [[UV1]], [[UV5]]
; F32DENORM: [[FADD2:%[0-9]+]]:_(s32) = G_FADD [[FMUL2]], [[UV9]]
; F32DENORM: [[FMUL3:%[0-9]+]]:_(s32) = G_FMUL [[UV]], [[UV4]]
; F32DENORM: [[FADD3:%[0-9]+]]:_(s32) = G_FADD [[FMUL3]], [[UV8]]
; F32FLUSH-LABEL: name: test_fmad_v4s32
; F32FLUSH: [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; F32FLUSH: [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
; F32FLUSH: [[COPY2:%[0-9]+]]:_(<4 x s32>) = COPY $vgpr8_vgpr9_vgpr10_vgpr11
; F32FLUSH: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<4 x s32>)
; F32FLUSH: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](<4 x s32>)
; F32FLUSH: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY2]](<4 x s32>)
; F32FLUSH: [[FMAD:%[0-9]+]]:_(s32) = G_FMAD [[UV]], [[UV4]], [[UV8]]
; F32FLUSH: [[FMAD1:%[0-9]+]]:_(s32) = G_FMAD [[UV1]], [[UV5]], [[UV9]]
; F32FLUSH: [[FMAD2:%[0-9]+]]:_(s32) = G_FMAD [[UV2]], [[UV6]], [[UV10]]
; F32FLUSH: [[FMAD3:%[0-9]+]]:_(s32) = G_FMAD [[UV3]], [[UV7]], [[UV11]]
; F32FLUSH: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[FMAD]](s32), [[FMAD1]](s32), [[FMAD2]](s32), [[FMAD3]](s32)
; F32FLUSH: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[BUILD_VECTOR]](<4 x s32>)
%0:_(<4 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
%1:_(<4 x s32>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
%2:_(<4 x s32>) = COPY $vgpr8_vgpr9_vgpr10_vgpr11
%3:_(<4 x s32>) = G_FMAD %0, %1, %2
$vgpr0_vgpr1_vgpr2_vgpr3 = COPY %3
...
|